使用paddleocr训练的DB++算法的文本检测模型，报错如下

[cola2023]

I0423 17:05:06.850246 10772 analysis_predictor.cc:714] ======= optimize end =======
I0423 17:05:06.851260 10772 naive_executor.cc:98] --- skip [feed], feed -> x
I0423 17:05:06.860247 10772 naive_executor.cc:98] --- skip [sigmoid_17.tmp_0], fetch -> fetch
I0423 17:05:06.872246 10772 naive_executor.cc:98] --- skip [feed], feed -> x
I0423 17:05:06.882247 10772 naive_executor.cc:98] --- skip [sigmoid_17.tmp_0], fetch -> fetch
Exception in thread "main" ai.djl.translate.TranslateException: java.lang.IllegalArgumentException: need at least one array to stack
at ai.djl.inference.Predictor.batchPredict(Predictor.java:187)
at ai.djl.inference.Predictor.predict(Predictor.java:124)
at me.aias.example.OcrV3DetectionExample.main(OcrV3DetectionExample.java:47)
Caused by: java.lang.IllegalArgumentException: need at least one array to stack
at ai.djl.util.Preconditions.checkArgument(Preconditions.java:31)
at ai.djl.ndarray.NDArrays.stack(NDArrays.java:1822)
at ai.djl.ndarray.NDArrays.stack(NDArrays.java:1784)
at me.aias.example.utils.detection.OCRDetectionTranslator.boxes_from_bitmap(OCRDetectionTranslator.java:277)
at me.aias.example.utils.detection.OCRDetectionTranslator.processOutput(OCRDetectionTranslator.java:110)
at me.aias.example.utils.detection.OCRDetectionTranslator.processOutput(OCRDetectionTranslator.java:27)
at ai.djl.inference.Predictor.batchPredict(Predictor.java:168)
... 2 more

Process finished with exit code 1

请问是不支持文本检测算法DB++吗？如果支持的话，请问上面这个错误是什么原因。如果不支持的话，请问是否有支持DB++算法的计划，如果有计划大概是什么时候，如果不准备支持，那么可以指导下从哪里下手改吗